The Impact of Training Programme Type and Duration on the Employment Chances of the Unemployed in Ireland

Seamus McGuiness, Philip O'Connell, Elish Kelly

    Research output: Contribution to journalArticlepeer-review

    12 Citations (Scopus)


    In the extensive literature on the employment impact of public sponsored training programmes for the unemployed, insufficient attention has been paid to the differential impact of different types of training programmes and of their varying duration. This paper uses a unique dataset, which tracks the labour market position of a cohort of unemployment benefit claimants for almost two years, to evaluate the impact of a range of government sponsored training courses in Ireland. Overall, we found that those who participated in training were less likely to be unemployed at the end of the two year study period. However, the average effect of training varied by the type and duration of training received. We found strong positive effects for job search skills training and medium to high level skills courses, a more modest positive effect for general vocational skills programmes (which are not strongly linked to demand in the labour market) and less consistent effects with respect to low level skills training. We also found that training episodes with lower duration had a more positive impact, with the exception of high level skills training programmes where longer training durations appear more effective. We ensure the robustness of our results by employing propensity score matching to reduce the impact of nonrandom assignment of programme participants, and estimate generalised propensity scores to estimate dose response functions.

    Original languageEnglish
    Pages (from-to)425-450
    Number of pages26
    JournalThe Economic and Social Review
    Issue number3
    Publication statusPublished - 2014


    Dive into the research topics of 'The Impact of Training Programme Type and Duration on the Employment Chances of the Unemployed in Ireland'. Together they form a unique fingerprint.

    Cite this